Cy3B™: Improving the Performance of Cyanine Dyes

P1: IZO

Journal of Fluorescence [JOFL] pp1140-jofl-481425 February 13, 2004 11:28 Style file version 29 Aug, 2003

Journal of Fluorescence, Vol. 14, No. 2, March 2004 (©2004)

Cy3BTM : Improving the Performance of Cyanine Dyes

Michael Cooper,1,3 Andreas Ebner,2 Mark Briggs, 1 Miles Burrows,1

Nicholas Gardner,1 Robert Richardson,1 and Richard West1

Received October 3, 2003; revised November 4, 2003; accepted November 5, 2003

The spectral properties of a rigidified trimethine cyanine dye, Cy3BTM have been characterised.This probe has excellent fluorescent properties, good water solubility and can be bioconjugated. Theemission properties of this fluorophore have also been investigated upon conjugation to an antibody.This study compared the conjugated emission properties of Cy3B with other commercially availablefluorophores emitting at similar wavelengths.

KEY WORDS: CyDye; Cy3B; bioconjugation; fluorescence.

INTRODUCTION

The use of technologies that rely on fluorescence de-tection is increasing in both the fields of drug discov-ery and high-throughput screening [1]. Techniques withsensitivity approaching that of radiolabelling, but withoutthe environmental concerns, are driving the need for ro-bust and reproducible assays formatted in a fluorescencemodality. Until recently, a limiting factor in the growth ofthis field lay also with the lack of suitable, commerciallyavailable instrumentation. This has been addressed withthe evolution of turnkey, platform technologies format-ted for bothin vitro [2] and cellular screening [3]. Today,such available platforms are capable of screening up to1 million compounds a day using a host of different fluo-rescence detection modalities [4]. Moreover, performinghigh volume screens also relies on the availability of ro-bust and sensitive fluorescence probes that are compatiblewith the biological nature of the assay. The current trendtowards assay miniaturisation [5] from 96via 384to 1536well plates requires highly soluble fluorophores with large

1 Amersham Biosciences, Discovery Systems, The Maynard Centre,Whitchurch, Cardiff, CF14 7YT, United Kingdom.

2 Institute of Biophysics, Johannes-Kepler University, AltenbergerStrasse 69, A-4040 Linz, Austria.

3 To whom correspondence should be addressed. E-mail: [email protected].

extinction coefficients, good photostability and quantumyields approaching unity. Furthermore, assays are oftenperformed in less than a 100µL of solution and at concen-trations that complement the nature of the bioassay thus,the fluorescent signal must be resolvable at sub-nanomolarconcentrations in both a rapid and reproducible manner.

Here, we describe the use of a high performance flu-orescent probe designed to meet the above criteria. Formore than a decade, the cyanine dye [6] (CydyeTM) rangeof fluorophores has been employed in a range of technolo-gies involved with discovery science. Wessendorfet al.[7] described Cy3 as the “archetypal” bright fluorophorein comparative studies with other commercially avail-able molecular probes including fluorescein, rhodaminederivatives and Texas red. More recently, Gruberet al.[8]have investigated the properties of Cy3 upon conjugationto macromolecules, such as immunogens, and the subse-quentanomalousincrease in emission relative to that of anequimolar solution of the free dye. Work described here,has therefore been undertaken to investigate the propertiesof a rigidified trimethine cyanine dye, Cy3B [9].

Nature, has a host of examples whereby trapping achromophore within a restrained system has maximisedthe fluorescence output. Allophycocyanin [10] (APC) isone such fluorophore whereby the conjugated system isheld within a proteinaceous matrix that maintains thechromophore in rigid state optimal for light absorptionand subsequent emission. Indeed, APC is available in a

1451053-0509/04/0300-0145/0C© 2004 Plenum Publishing Corporation

P1: IZO


146 Cooper, Ebner, Briggs, Burrows, Gardner, Richardson, and West

Fig. 1. The structures of Cy3 (1) and Cy3B (2).

conjugatable format (XL665) for screening purposes, al-though the high molecular weight of>100 kDa can bea hindrance in some assay formats [11] and, due to theproteinaceous nature of the probe it can suffer from longterm stability issues.

A similar approach has been taken to optimise thespectral properties of Cy3B. This probe has the same com-mon chromophore as Cy3 (Fig. 1). Open chain trimethinecyanine dyes e.g. Cy3 can exist in many conformationsand are capable of undergoing numerous rotational andtranslational modes of vibration which result in the lossof energy by non-radiative processes and a resultant low-ering ofφ [12]. In Cy3B, a rigid backbone has thereforebeen introduced to provide a trimethine dye with a fixedconformation in both the relaxed and excited states. Thisis in a bid to increase the efficiency of the fluorescent out-put upon excitation. Within the design criteria, essentialfunctionalities such as water solubility (via sulfonation),maintaining a low molecular weight and the ability to bio-conjugate (viaN-hydroxy succinimidyl ester) have beenconserved.

EXPERIMENTAL

The synthetic procedures for the synthesis of Cy3Bare described elsewhere [9]. PD10 gel filtration columnsand Cy3B and Cy3 were obtained from Amersham Biso-ciences. 5-TAMRA, Texas Red, BODIPY-TMR-X andAlexa 555 were obtained from Molecular Probes, Eugene,OH, Innosense (now Cyanine Technologies) IRIS-3 andInnosense IRIS fuchsia were obtained from Web ScientificLtd, Crewe, UK and Atto-Tec Atto 565 was obtained fromFluka, Dorset, UK. All dyes were purchased asN-hydroxysuccinimidyl esters. Goat IgG was obtained from SigmaAldrich, Dorset, UK.

All spectroscopic measurements were made viawavelength scans. UV-vis spectrophotometry was carriedout using a Hitachi U-3310 with either a deuterium lamp(200–340 nm) or tungsten iodide lamp (340–800 nm) inabsorbance mode. The scan speed was 120 or 300 nm/minwith a slit width of 2 nm.

Fluorescence spectrophotometry was carried out us-ing a Hitachi F-4500 with 150 W Xenon lamp in emissionmode. The scan speed was 60 or 240 nm/min with an ex-citation slit width of 5 nm and an emission slit width of5 nm. The PMT voltage was 700 V.

Buffer Solutions

Buffer 8.3 (coupling buffer): 100 mM NaCl, 35 mMH3BO3, pH 8.3 (with NaOH)

Buffer 7.5(= buffer A): 100 mM NaCl, 50 mM NaH2PO4,1 mM EDTA, pH 7.5 (with NaOH)

Buffer 9: 100 mM NaCl, 50 mM NaH2PO4, 1 mM EDTA,pH 9 (with NaOH)

Stop buffer(= stop solution): 100 mM NaH2PO4·H2O

IgG Stock Solution

1 mg goat IgG/mL coupling buffer, pH 8.3.Stock solutions were prepared freshly by dilution of

goat IgG with buffer 8.3. Concentrations for exact dilutionand final determination were measured by UV-vis spec-troscopy: [IgG]= (A280− kcorr dye× Adye max)/210.000.

Dye Stock Solutions

The most concentrated stock solution of each dye“stock solution A” was prepared by dissolving a defined

P1: IZO


Improving the Performance of Cyanine Dyes 147

mass of the solid dye in a defined volume of DMF (e.g.2 portions of Cy5 sold as sufficient “to label 1 mg ofprotein” were diluted with 100µL DMF), e.g. 50µLdye stock A was diluted with 50µL DMSO to give100 µL stock solution B, and so on. All further dyestock solutions were prepared by serial dilution withDMSO to give a range of solutions of different dyeconcentration designed to facilitate labelling at differentdye:antibody stoichiometries. The concentration of dyein the most concentrated dye stock solution (A) was cal-culated from the UV-vis spectrum using the followingequation.

[dye]= (Amax dye/εye)

In particular, 8µL of the DMF stock solutions were dilutedwith 1 mL of a 10 mM Na2CO3 solution to hydrolyse thedye. After 10 min this solution was diluted with 25 mL ofbuffer 7.5 and the spectrum was measured. The absorptionat 280 nm was divided by the absorbance at the absorp-tion maximum of the dye to give the 280 nm correctioncoefficient (kcorr dye) for each dye.

Standard Procedure for Goat IgG Labellingwith Fluorescent Probe

10µL of a particular dye stock solutions was addedto 0.2 mL buffer 8.3 containing 0.2 mg goat-IgG whilevortexing and subsequently left to react for 60 min in thedark. This was further vortexed every 10 min during thisreaction period. The reaction was terminated with 300µLof stop buffer.

Separation of labelled IgG from free dye was carriedout by gel filtration on Amersham PD-10 columns. Thecolumn was pre-washed with 20 mL buffer 7.5, 500µLof the labelling solution was loaded and allowed to drain.Then, 1 mL buffer 7.5 was added and allowed to drain(2x). For elution of the labelled IgG, 2 mL of water wasadded and the eluate collected. Finally, 10 mL water wasadded to elute the unbound dye.

UV-Vis Measurements

The number of bound dyes per antibody was mea-sured by UV-vis spectroscopy. The antibody concentra-tion was measured by absorption at 280 nm corrected forthe dye absorption at 280 nm. [IgG]= (A280− kcorr dye×Adye max)/210.000.

The determination of the correction coefficient is asdescribed above.

The dye concentration was also measured: [dye]=(Amax dye/εdye).

Table I. Spectral Properties of Cy3 and Cy3B

λ max.absλ max.em.ε (L·mol−1 RelativeCompound (nm) (nm) cm−1) τ (ns) φ brightness

Cy3 550 570 150,000<0.3 0.04 1Cy3B 558 572 130,000 2.8a 0.67a 7.9

aNon Radiative lifetime= 3.19 ns,φf = 0.88.

Fluorescent Measurements

Each solution of labelled IgG, and free dye, was di-luted to dye concentration of 50 nM with buffer pH 7.5.For each sample, the fluorescence spectra were recordedat different excitation wavelengths, at 5 nm intervals. Thelongest wavelength of excitation was close to the wave-length maximal absorption. Non radiative lifetime andquantum yield were calculated by spectral means [13] us-ing trapezoidal numerical integration.

Spectral Properties of Cy3B

The spectral properties of Cy3B have some com-monalities with Cy3. These are summarised in Table I.The rigidification has the effect of shifting the absorptionspectrum of Cy3B by several nanometers (Fig. 2). More-over, the pronounced effect of rigidification can be seenin the emission spectrum. An∼8 fold increase in fluores-cence was observed with an equimolar solution of Cy3Bwhen compared to Cy3. This is illustrated in Fig. 3. UsingRhodamine 6G as a standard (φf = 0.94 [14]), the flu-orescence quantum yields for both Cy3 and Cy3B werealso measured and found to have aφf for Cy3B of∼0.67compared to 0.04 for Cy3 in aqueous solution.

Fig. 2. Absorption spectra of Cy3 and Cy3B.

P1: IZO



Fig. 3. Emission spectra of equal concentrations of Cy3 and Cy3B(excitation 550 nm).

The lifetime properties of Cy3B are also modifiedcompared to Cy3. The rigid structure of Cy3B has theeffect of increasing the mean fluorescent lifetime of theprobe to∼2.9 ns (non-radiative lifetime= ∼3.1 ns).

Bioconjugate Properties of Cy3B

The work of Gruberet al. [8] has recently illustratedthe importance of measuring the effectiveness of fluores-cent probes upon bioconjugation. Cy3B, like other Cydyereagents is, by design, a fluorescent probe for biologicallabelling and therefore any post-labelling change in theemission properties needs to be well understood. This isespecially true when the bioconjugate is to be employedfor the quantification of a biomolecular event.

A study was made of the emission properties for Cy3and Cy3B, in comparison with a range of other spectrallymatched, commercially available dyes. This was carriedout after bio-conjugation of the probes to goat IgG andat a range of dye-to-antibody concentrations. A numberof stock solutions of the different dyes were prepared thatallowed a standard concentration of goat IgG to be labelledwith different dye stoichiometries. After purification thesesolutions provided a range of known dye:antibodies ratioscalculated from the UV-vis absorption spectra.

For the purpose of these experiments, relative bright-ness is defined as the fluorescent output of a range dye-labelled antibodies as compared using fluorescence spec-troscopy. The level of dye labelling has been calculatedusing the absorption properties of the fluorescent probein each case. An important assumption made during thisstudy is that the molar absorptivity of the dyes doesnot change upon conjugation. The generation of absolute

dye:protein data was considered to be beyond the scopeof this study. Precedent [8] for this assumption is avail-able and it has demonstrated that spectroscopic methodscan provide important information regarding the trendsof dye-labelled conjugates. Therefore, the findings of thispublication demonstrate important methodologies for at-taining relative comparable data. Such techniques offer aninsight into the behaviour of these fluorescent probes uponbioconjugation.

RESULTS AND DISCUSSION

The data for the purified dye-antibody conjugates areshown in Fig. 4. A range of stoichiometries for each dye-antibody has been obtained and the emission characteris-tics of each solution have been measured. Figure 4 com-pares the emission characteristics of each dye at differentIgG labelling ratios.

From these results it can be seen that the fluores-cent performance of the different conjugated dyes is wideranging. Many of the probes display emission quenchedcharacteristics upon increased labelling. This observationis thought to be due to either F¨orster resonance energytransfer between like molecules, e.g. homotransfer [15] oralternatively, the formation of excimers. Homotransfer isa phenomenon made possible by the “anti-Stokes” overlapof the fluorescence emission spectrum and the absorptionspectrum oflike fluorophores located within a finite dis-tance. This phenomenon becomes more prevalent wherethe donor concentration is suitably high, such as, at suf-ficiently high dye-antibody ratios. It is thought that suchquenching can also expedited by the close association oraggregation oflike dyes in solution, subsequently leadingto proximal labelling events on the antibody or protein.Planar dyes or dyes with a propensity forπ -stacking aremore likely to exhibit such aggregation. Although, ex-cimer formation may offer a more plausible explaination.Here quenching occurs due to the formation of excitedstate dimers. Homotransfer is more likely to occur in amore rigid protein system than the model IgG system cho-sen. To measure the degree of quenching occurring forCy3B upon multiple labelling, the emission properties ofthe free dye were compared to the emission from the dif-ferent dye:protein ratios obtained for Cy3B labelled IgG.This is illustrated in Fig. 5. Here it can be seen that asthe dye:protein ratio increases, the subsequent quenchingincreases also. The absorption spectrum of the free dyecompared to that of a biomolecule labelled with>4 Cy3Bmolecules also observes an increase in the 520 nm shoul-der. This change in the spectral properties of the dye ismost likely due to some excimeric formation. Changes in

P1: IZO


Improving the Performance of Cyanine Dyes 149

Fig. 4. Comparative emission properties of dye-labelled goat IgG conjugates obtained from a range of commer-cially available fluorophore NHS esters.

molar absorptivity may also arise due to charged residuesprevalent upon the protein. These may have an effect uponthe fluorescence quantum yield [19] for the labelled probesif the pKa for such probes is close to that of the buffer usedin the experiment. In this instance, Cy3B (and Cy3) areknown to be insensitive over the physiological pH range.4

Some dyes exhibit an increase in emission upon pro-tein labelling. Gruber and co-workers have previously de-scribed this phenomenon for Cy3 [8]. This is often dueto the dye being maintained in a more rigid dye confor-mation upon biomolecular labelling than in the free statein solution. The lipophilic nature of the biomolecule mayalso shield the dye from solvent quenching. It is appar-ent from the data presented here, that compared to Cy3B,Cy3 is brighter upon biomolecular labelling than it is inthe free form. The ensemble experiments described ear-lier observed a∼8 fold increase in brightness for Cy3B infree form when compared to Cy3. The IgG labelling datashows a difference in brightness of∼2.0–2.5 for Cy3Bcompared to Cy3 at pertinent labelling concentrations.This is in agreement with a previous study [8], wherebyit was demonstrated that the quantum yield of Cy3 in-creases upon biolabelling of antibodies. Due to the speci-

4 Amersham Biosciences Unpublished data observe an increase in theshoulder at 520 nm upon the labelling streptavidin, for example with>4 molecules of Cy3B. This increase is indicative of excimeric dyeformation.

ficity of an antibody for an antigen, it is important thatany modification of such immunogens with fluorescentdyes is performed in the least intrusive and non-destructivemanner possible. This decreases the likelihood of reducedspecificity of the dye-labelled adduct. As such, it can beseen that the maximum fluorescence, for Cy3B, can be ob-tained from a Cy3B-antibody adduct with approximately3 dyes attached. Due to the high extinction coefficientand quantum yield of this probe, it can be seen that atall dye-to-antibody concentrations, this probe offers the

Fig. 5. Comparing the emission of the free Cy3B dye in solution withthe emission from variable dye: Protein concentrations.

P1: IZO



most efficient emission properties. Upon labelling the an-tibody with 3 Cy3B dyes, the fluorescence characteristicsare more than 2 fold that of Cy3 or Alexa 555 and greaterthan 3 fold for the remaining dyes employed in this study.Indeed, to obtain an equivalent fluorescence output fromthe other probes, up to 7 Alexa 555 dye labels are requiredor 10 Cy3 dye labels. At such high labelling concentra-tions, there is an increased probability that the specificityof the biological molecule will be affected [17]. It can beseen also that there is no increase in fluorescence for Cy3Bupon employing more than 3 dyes per antibody. It is at thisdye to protein concentration that some quenching can beobserved.

The data in Fig. 4 show that at comparatively lowlabelling ratios, Cy3B offers the most efficient fluorescentoutput. This permits biomolecular labelling using Cy3Bthat is relatively unintrusive and with a low risk of dam-aging the specificity of the molecule.

Cy3B also has a fluorescent lifetime of 2.9 ns(Table I). This offers Cy3B as a promising dye for flu-orescence polarisation. Recent work by Turconiet al.[18] described the use of Cy3B in a fluorescence polar-isation screening assay whereby higher anisotropy mea-surements were obtained with Cy3B in comparison withother probes. Biological screening using fluorescence po-larisation often requires sub-nanomolar concentrations offluorophore in order to complement the biological inter-actions under scrutiny. It is here that brighter and moresensitive red-shifted probes such as Cy3B are beginningto realise potential [19].

ACKNOWLEDGMENTS

The authors would like to thank Dr Rudi Labarbefor assistance with theoretical calculations and the Insti-tute of Biophysics, J. Kepler University, Linz for helpfuldiscussion.

REFERENCES

1. A. J. Pope, H. M. Ulrich, and K. J. Moore (1999). Homogeneousfluorescence readouts for miniaturized high-throughput screening:Theory and practice.Drug Discov. Today4(8), 350–362.

2. P. Ramm (1999). Imaging systems in assay screening.Drug Discov.Today4(9), 401–410.

3. P. Ramm and N. Thomas (2003). Technological advances in high-throughput screening. Image-based screening of signal transductionassays.Sci. STKE 2003, p. 14.

4. J. Masino, S. G. Jones, and J. A. Alcock (2002). inPoster Presentedat IBC’s 7th Annual Drug Discovery TechnologyTM World Congress,Boston, USA, 2002.

5. T. K. Garyantes (2002). 1536-well assay plates: When do they makesense?Drug Discov. Today7(9), 489–490.

6. R. B. Mujumdar, L. A. Ernst, S. R. Mujumdar, C. J. Lewis, and A. S.Waggoner (1993). Cyanine dye labelling reagents: Sulfoindocyaninesuccinimidyl esters.Bioconjug. Chem.4, 105–111.

7. M. W. Wessendorf and T. C. Brelje (1992). Which fluor is the bright-est. A comparison of staining obtained using fluorescein, TMR, rho-damine, Texas Red and Cy3.Histochemistry98, 81–85.

8. H. J. Gruber, C. D. Hahn, G. Kada, C. K. Riener, G. S. Harms,W. Ahrer, T. Dax, and G. H. G. Knaus (2000). Anomalous fluores-cence enhancement of Cy3 and Cy3.5 versus anomalous fluorescenceloss of Cy5 and Cy7 upon covalent linking to IgG and non-covalentbinding to Avidin.Bioconjug. Chem.11(5), 696–704.

9. A. S. Waggoner and R. Mujumdar (2001).Rigidized Trimethine Cya-nine Dyes, U.S. Patent 6133445A.

10. V. Oi, A. N. Glazer, and L. Stryer (1982). Fluorescent phycobilipro-tein conjugates for analyses of cells and molecules.J. Cell Biol.93,981–986.

11. E. Trinquet, F. Maurin, H. Bazin, and G. Mathis. (2003). inPosterPresentation at MAF Conference, Prague August 24th–27th, 2003.

12. D. F. O’Brien, T. M. Kelly, and L. F. Costa (1974). Excited state Prop-erties of some Carbocyanine dyes and the energy transfer mechanismof spectral sensitisation.Photogr. Sci. Eng. 18(1), 76–84.

13. J. R. Lakowicz (1999).Principles of Fluorescence Spectroscopy,2nd ed., Kluwer Academic, New York, p. 10.

14. M. Fischer and J. Georges (1996). Fluorescence quantum yield ofrhodamine 6G in ethanol as a function of concentration using thermallens spectrometry.Chem. Phys. Lett.260, 115–118.

15. B. Wieb Van Der Meer, G. Coker III, and S.-Y. Simon Chen (1994).Resonance Energy Transfer, Theory and Data, VCH publishers, NewYork.

16. R. Sjoback, J. Nygren, and M. Kubista (1998). Characterizationof fluorescein-oligonucleotide conjugates and measurement of lo-cal electrostatic potential.Biopolymers46, 445–453.

17. H. Karsilayan, I. Hemmila, H. Takalo, A. Toivonen, K. Pettersson,T. Lovgren, and V.-M. Mukkala (1997). Influence of couplingmethod on the luminescence properties, coupling efficiency, andbinding affinity of antibodies labeled with Europium(III) chelates.Bioconjug. Chem.8(1), 71–75.

18. S. Turconi, K. Shea, S. Ashman, K. Fantom, D. L. Earnshaw, R. P.Bingham, U. M. Haupts, M. J. B. Brown, and A. J. Pope (2001).Real experiences of uHTS: A prototypic 1536-well fluorescenceanisotropy-based uHTS screen and application of well-level qual-ity control procedures.J. Biol Screen. 6(5), 275–290.

19. T. C. Turek-Etienne, E. C. Small, S. C. Soh, T. A. Xin, P. V. Gaitonde,E. B. Barrabee, R. F. Hart, and R. W. Bryant (2003). Evaluation offluorescent compound interference in 4 fluorescence polarizationassays: 2 Kinases, 1 Protease, and 1 Phosphatase.J. Biol. Screen.8(2), 176–184.

Documents

Cy3B™: Improving the Performance of Cyanine Dyes